There must be a million daft little daily quizzes online, so one more couldn’t hurt? Building on the principles behind my tmdb.org/Bluesky bot (and my mild obsession with cinema), I created a simple Wordle-style game: Ten questions per day, name the actor/director/composer.
As is the fashion of today, of course it tracks your scores and streak (locally, nothing is stored on the server).
In order to give players half a chance, it grabs only films above a certain rating and vote count on tmdb. I say certain, because I had to tier the scores and counts by decade, users of tmdb have understandably been much more diligent scoring newer films, so anything older than the 1980s kept being crowded out.

Some other peculiar things popped up that needing sorting, mainly because of the actual history of cinema. I had to address the gender imbalance by both including supporting actors as decoys (there are far more male stars, and especially directors and composers, who knew) and enforcing at least 40% women in the questions. The decoys are also all matched by gender and birth era to the correct answer, you don’t want Michael J. Fox as an option when you’re guessing who played Charles Foster Kane.

Oh, and there’s AI! Sort of. A local Ollama instance (currently using gpt-oss-20b, but that might well change) runs a weekly analysis on question difficulty, era spread and gender balance. This doesn’t really do much at the moment, but once it’s amassed enough data on questions/scores I *might* let it have a go at setting questions.
Currently there’s a playable archive of every past game, I’ll probably limit this to the past month once it gets unwieldy.

Analytics are just using a self-hosted Umami instance, so it’s cookieless and wonderfully GDPR-compliant.

If you’ve read this far you should probably go outside for a bit, right after you’ve tried the game at classicfilms.app.